Single-cell transcriptomics reveal individual and cooperative effects of trisomy 21 and GATA1s on hematopoiesis

Abstract

Trisomy 21 (T21) is associated with baseline erythrocytosis, thrombocytopenia, neutrophilia, transient abnormal myelopoiesis (TAM), and myeloid leukemia of Down syndrome (ML-DS). TAM and ML-DS blasts harbor mutations in GATA1, resulting in the exclusive expression of the truncated isoform GATA1s. Germline GATA1s mutations in individuals without T21 cause congenital cytopenias, typically without a leukemic predisposition. To dissect the developmental effects of T21 and GATA1s, we used a combination of isogenic human induced pluripotent stem cells, primary human fetal and neonatal cells, and single-cell transcriptomics to interrogate hematopoietic progenitors differing only by chromosome 21 and/or GATA1 status. Both T21 and GATA1s induced early lineage skewing, and trajectory analysis revealed that GATA1s altered the temporal regulation of lineage-specific transcriptional programs, disrupting cell proliferation and maturation irrespective of chromosomal context. These studies uncovered unexpected heterogeneity and lineage priming in early, multipotent hematopoietic progenitors and identified transcriptional and functional maturation blocks linked to GATA1s.

Keywords: Down syndrome; GATA1; GATA1s; Trisomy 21; hematopoiesis; hematopoietic progenitors; iPSC; lineage priming; single-cell RNA sequencing; transient abnormal myelopoiesis.

Graphical abstract

Figure 1

Characterization of hematopoietic cells differentiated from isogenic iPSCs (A) Western blot of GATA1 and GATA1s expression in iPSC-derived MKs. (B) EB-based hematopoietic differentiation protocol. (C) Flow cytometry plots of day 7, 9, and 11 HPCs differing only by HSA21 and/or GATA1 status. CD41⁺/235⁺: multipotent, CD41⁺/235⁻: MK bias, CD41⁻/235⁺: erythroid bias, CD41⁻/235⁻: myeloid bias. (D) CD41⁺/42b⁺ MK fold change of iPSC-derived HPCs differentiated in MK liquid culture for 6 days. (E) CD235⁺/71⁺ erythroid fold change of iPSC-derived HPCs differentiated in erythroid liquid culture for 12 days. For (D) and (E), data represented as mean ± SEM; n = 8–17 independent experiments per clone. ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗∗p < 0.0001; ns, not significant. Also see Figure S1.

Figure 2

scRNA-seq characterization of primary and iPSC-derived HPCs (A) Isolation of fetal liver HPCs, TAM blasts, and CD41⁺/235⁺ iPSC-derived day 7 HPCs. (B) Annotated integrated UMAP of cells from (A). Also see Figures S2 and S4. (C) Individual annotated UMAPs for each sample. (D–F) Relative expression of hematopoietic and cell-cycle genes in (D) erythroid-, (E) MK-, and (F) myeloid-biased clusters, organized by genotype and sample type. Also see Figures S3A–S3C.

Figure 3

scRNA-seq characterization of HPCs generated from isogenic iPSCs differing by HSA21 and/or GATA1 status (A) Annotated integrated UMAP of euploid/wtGATA1, euploid/GATA1s, T21/wtGATA1, and T21/GATA1s iPSC-derived HPCs collected on days 7, 9, and 11 of differentiation. (B) Relative expression of hematopoietic genes, organized by lineage/cluster. Also see Figure S5. (C) Individual annotated UMAPs for each genotype and time point. ^∗indicates distinct population of MK-committed HPCs in euploid/wtGATA1 cells. Also see Figures S3D and S3E. (D) Relative proportions of lineage-skewed HPCs for each genotype and time point.

Figure 4

Differentially expressed hematopoietic genes and enriched pathways in isogenic iPSC-derived HPCs (A–D) Relative expression of lineage markers in (A) unbiased HPC, (B) erythroid-, (C) MK-, and (D) myeloid-biased clusters organized by genotype and time point. Also see Figures S3F–S3I. (E–H) Selected GSEA of (E) day 7 and (F) day 11 euploid cells and (G) day 7 and (H) day 11 T21 cells, organized by cluster. NES, normalized enrichment score.

Figure 5

Trajectory analysis of isogenic iPSC-derived HPCs (A) Projection of latent time indicating degree of differentiation and lineage maturation. 0 = least mature and 1 = most mature. (B) Terminal states inferred by latent time analysis. (C) Inferred mean trajectories for MK, erythroid, and myeloid cells across all time points, separated by genotype. Vertical line indicates median; left shift indicates less mature. (D and F) Heatmap showing developmental expression along latent time for (D) MK and (F) erythroid driver genes. Genes ordered by peak expression time as cells progress toward terminal state. (E and G) GSEA of (E) MK and (G) erythroid driver genes. Also see Figure S6.

Figure 6

T21 and GATA1s impair maturation of MKs and erythroid cells (A) Composite CD41 median fluorescence intensity (MFI) of MKs. n = 5–8 independent experiments per genotype. (B) Composite flow cytometric analyses for PAC-1 binding after thrombin stimulation of MKs. n = 4–6 independent experiments per genotype. (C) Composite CD42b MFI of MKs. n = 4–10 independent experiments per genotype. (D) Composite flow cytometric analyses for PAC-1 binding after thrombin stimulation of MKs derived from iPSCs of a separate genetic background. n = 5–6 independent experiments per genotype. (E) Composite CD42b MFI of MKs derived from iPSCs of a separate genetic background. n = 5–8 independent experiments per genotype. (F) Erythroid cell morphology assessed at 10× magnification. (G) Composite flow cytometric analyses of Band3 expression of erythroid cells. n = 3 independent experiments per genotype. For (A–E) and (G), MKs were assayed on day 6 and erythroid cells were assayed on day 12 of single-lineage liquid cultures. Data represented as mean ± SEM.